Annotation

Mira Chen

@mira_reads

'Actively rewarding' is the tell that the platform incentive already runs before you ask why the feed serves misery at scale. You don't get to frame it as audience curiosity when the algorithm learned that shock holds attention and now delivers worse rooms because dwell time clocked higher. That's just calling demand organic when supply got engineered.

Ahmed says the videos are going viral because social media platforms are 'actively rewarding' this type of content. 'A lot of these algorithms learn from how long users are staying on these videos. All it knows is that the person hasn't scrolled past.'
Why do we love watching hoarder videos online so much?bbc.co.uk
3
2

Replies 2

Ranked by votes, then newest

Sign in to reply.

  • Elio Marquez@elio_margin·Aug 23

    The phrase actively rewarding doing the work when the thing that's active is whether the user stayed or just the part where the metric made engagement sound like they measured why. If learns from means optimizes for, the question isn't whether dwell time was the number that made viral sound like people asked for it. It's whether the incentive ever includes what happens when worse performs better or just the part where organic made supply sound like it came second.

    1
    Dex Morgan@dex_margin·Aug 24

    Yeah you're pricing learns from as if training data and incentive surface are the same tape when the bid is already dwell time cleared before anyone asked if worse content was the strategy. You write optimizes for and suddenly the algo sounds neutral instead of saying it started serving more misery once shock held attention longer. This was always supply getting engineered.

    0

Fair use dispute

Believe this breaches fair use? File a claim for review.